vascular calcification is widespread in patients with coronary artery disease and peripheral artery disease (21) and is closely associated with the incidence of cardiovascular events as well as all-cause mortality (3, 27, 34). Calcification in the tunica media is often observed in elderly people and is highly correlated with their morbidity and mortality (8).

Many recent findings have suggested that vascular calcification is regulated by the machinery similar to bone formation, which is accomplished through the extracellular matrix (ECM) calcification (16, 21, 34). During the ECM calcification, hydroxyapatite crystals that contain calcium and inorganic phosphate precipitate within the collagen fibrils (32). Many key players in the ECM calcification, such as matrix Gla protein (MGP) and alkaline phosphatase (ALP), have been identified (22, 32). Inorganic pyrophosphate, a small molecule made of two phosphate ions, and MGP prevent incorporation of mineral crystals into the collagen fibrils (34). In contrast, ALP promotes ECM calcification by cleaving pyrophosphate (32). Osteoblasts play a central role in the ECM calcification by producing both ALP and type I collagen, the major component of the ECM in bone matrix.

These bone calcification regulatory factors have been identified in blood vessels, particularly at sites of medial calcification and calcified atherosclerotic plaques (9, 11, 40, 45). Moreover, expression of these factors is differentially regulated between nondiseased and diseased vessels. In the vessels of senescence-associated medial calcification, MGP mRNA expression was reduced, whereas ALP mRNA expression was enhanced, compared with that in normal vessels (40). Calcification in the media occurred in the absence of macrophages and lipid and was associated with α-smooth muscle actin-positive vascular smooth muscle cells (VSMCs), suggesting the possible role of senescent VSMCs in the formation of medial calcification (40). However, it remains unclear what induces the altered expression of the MGP and ALP in the calcified vessels and how VSMCs are involved in the formation of senescence-associated medial calcification.

As the population ages, senescence-associated medial calcification is becoming a progressively more common and important problem. However, little is known about the molecular mechanisms governing the medial calcification. Involvement of cellular senescence in VSMCs in the formation of medial calcification is still unclear. In addition, it remains to be determined whether pharmacological intervention can prevent the VSMC calcification that occurs in the senescent cells.

In the present study, we examined the potential role of replicative senescent VSMCs in the formation of senescence-associated medial calcification. We found that replicative senescence of VSMCs enhanced the susceptibility to calcification by increasing the expression of osteoblastic genes. Furthermore, we demonstrated that statin and Rho-kinase inhibitor effectively inhibited the calcification in the senescent VSMCs.

MATERIALS AND METHODS

Materials.

Antibody for smooth muscle α-actin (α-SMA) was obtained from Sigma (St. Louis, MO). Antibodies for p16^Ink4a and p21^Waf1/Cip1 were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies for SM22α and SM-myosin heavy chain (MHC)-2 were obtained from Abcam (Cambridge, UK). Fluvastatin was kindly provided by Novartis. Human coronary artery smooth muscle cells were obtained from Cambrex (Charles City, IA).

Mice.

Klotho-deficient mice were obtained from Clea Japan. All animal experiments were reviewed and approved by the Institutional Animal Care and Use Committee of the Kyoto Prefectural University of Medicine.

Cell culture.

VSMCs were cultured in SmGM-2 medium (Cambrex). When cells reached confluence, 20% of the cells were replated and cultured until they again became confluent for the next passage. Senescence-associated β-galactosidase activity was analyzed as previously described (10). To confirm the consistency of the results, we used two lines of human primary VSMCs from a distinctive lot. One line of VSMCs reached passage 13, whereas the other reached passage 11. VSMCs in early passage (passage 6) were used as control young cells. Comparison between the senescent and the control young VSMCs was always performed by using the cells from the identical line. For the in vitro calcification assay, VSMCs were cultured in the calcification medium (DMEM supplemented with 15% FBS in the presence of 2.5 mM phosphate). To investigate the effect of fluvastatin or Y-27632 on gene expression and phosphorylation of Smad 1/5/8, we cultured VSMCs in the basal SmBM medium supplemented with 5% FBS. For other experiments, VSMCs were always cultured in the regular growth medium (SmGM-2).

Quantitative PCR.

Total RNA was isolated from cells using Trizol (Invitrogen, Carlsbad, CA), followed by cDNA synthesis using the First-Strand cDNA synthesis kit (Invitrogen). Quantitative PCR was performed by LightCycler (Roche Applied Science, Auckland, New Zealand) using the FastStart DNA Master Plus SYBR green I kit (Roche Applied Science). Expression levels of target genes were all corrected using GAPDH expression. Primers used are shown in Table 1. The PCR conditions were 95°C for 12 s, 57°C for 12 s, and 72°C for 15 s.

Immunoblotting, immunocytochemistry, and immunohistochemistry.

Immunoblotting, immunocytochemistry, and immunohistochemistry were performed as previously described with minor modifications (18, 19). Briefly, cells were lysed in RIPA buffer containing protease inhibitor cocktail (Sigma) and phosphatase inhibitors. After measurement of protein concentration using the DC protein assay kit (Bio-Rad Laboratories, Hercules, CA), the same amount of crude cell lysates were run on 15% SDS-PAGE gel, followed by immunoblotting. For immunocytochemistry, cells were fixed with 4% paraformaldehyde and then permeabilized with 0.1% Triton X-100. Cells were then incubated with anti-SM22α antibody, anti-α-smooth muscle actin (SMA) antibody, or anti-SM-MHC2 antibody at room temperature for 1 h, followed by detection using fluorescence-labeled secondary antibodies (Invitrogen). For immunohistochemistry, mouse aorta was fixed with 4% paraformaldehyde, followed by paraffin embedding. After being blocked with methanol and 10% normal rabbit serum, sections were incubated with the first antibodies at 4°C overnight. The same concentration of normal goat IgG was used as a negative control. Signals were visualized using the DAB enzyme kit (Nichirei, Japan).

Measurement of ALP activity.

Cellular ALP activity in the cells was visualized by cytochemical staining. After being washed with PBS, cells were fixed with citrate-acetone-formaldehyde fixative solution for 15 min. The cells were then washed twice with deionized water and stained for 15 min with an alkaline-dye mixture containing sodium nitrite, FRV-alkaline, and naphthol AS-BI alkaline solutions (Sigma).

Quantitative analysis of ALP activity was performed using the SensoLyte pNPP ALP assay kit (AnaSpec, San Jose, CA). Briefly, cells were incubated for 30 min in AP reaction buffer (0.5 ml of 0.75 M 2-amino-2-methyl-1-propanol with 0.5 ml of 2 mg/ml p-nitrophenyl phosphate). The reaction mixture was then mixed with 0.1 N NaOH, followed by measurement of optical density at 405 nm.

In vitro VSMC calcification.

VSMCs were incubated in the calcification medium (DMEM supplemented with 15% FBS in the presence of 2.5 mM phosphate) for 4–7 days. Medium was changed every other day. At the end of incubation, cells were fixed with 70% ethanol, followed by staining with 2% Alizarin red-S (pH 4.2) for 5 min. Mineralized area was measured in three to five independent wells for quantification. Also, precipitated calcium was quantified by the colorimetric analysis of calcium content. Briefly, cells were decalcified by incubation in 0.6 M HCl, and the calcium content was measured using the Calcium E-test kit (Wako, Osaka, Japan). Remaining cells were lysed in 0.1 M NaOH-0.1% SDS, and the protein concentration was measured using the DC protein assay kit (Bio-Rad). Terminal deoxynucleotidyl transferase dUTP-mediated nick-end labeling (TUNEL) staining of cells incubated in the calcification medium was performed using an in situ cell death detection kit (Roche Applied Science). In some experiments, VSMCs were incubated in the calcification medium in the presence of fluvastatin or Y-27632 (Rho-kinase inhibitor) with or without warfarin. Cells were regularly given fresh medium every other day. To examine the effect of statin or Y-27632 on VSMCs apoptosis, cells were incubated in the serum-free calcification medium as previously described (41), followed by the TUNEL staining.

Knockdown of ALP, type I collagen, and runt-related transcription factor-2.

Short interfering RNAs (siRNA) for ALP, type I collagen, and runt-related transcription factor-2 (RUNX-2) were all obtained from Dharmacon (Lafayette, CO). Negative control siRNA (Ambion, Austin, TX) was used as a control (scramble). Short interfering RNAs were transfected into cells using RNAiMax (Invitrogen) as the manufacturers recommended. Effective transfection of siRNA and the knockdown of the target genes were confirmed using Cy5-labeled negative control siRNA (Ambion) (data not shown) and quantitative PCR, respectively. Cells were given fresh growth medium 24 h after transfection, and total RNA was isolated from cells 36 h after transfection. For the VSMC calcification assay, cells of the same number were replated into 96-well plates 48 h after transfection, followed by incubation in the calcification medium for 4 days.

Statistical analysis.

Differences between groups were analyzed using Student's t-test, with P < 0.05 considered significant. Data are means ± SE as indicated.

RESULTS

Generation of senescent VSMCs.

Normal somatic cells enter a state of irreversible growth arrest and altered function after a finite number of divisions. This process is termed replicative senescence and is thought to be a tumor-suppressive mechanism and an underlying cause of aging (10). Because age-related loss of proliferative activity in human VSMCs has been reported (37), we used the replicative senescent VSMCs as an in vitro model to study the role of senescent VSMCs in the vascular calcification. To induce the replicative senescence, we cultured primary VSMCs from human coronary artery for an extended period until the proliferative activity was lost. We used VSMCs in early passage as control young cells. Senescent VSMCs demonstrated flat and enlarged morphology as well as significant senescence-associated β-galactosidase activity (10) (Fig. 1A). Furthermore, expression of p16^Ink4a and p21^Waf1/Cip1, known as senescence-related genes (2, 24, 29), was upregulated in the senescent VSMCs (Fig. 1B). On the other hand, significant expression of SM22-α, α-SM actin, and SM-MHC2 was detected in the senescent VSMCs at the level similar to the control young cells, excluding the possibility that other types of cells dominated the cell population during the cell culture for an extended period (Fig. 1C).

Senescence of VSMCs enhances the susceptibility to calcification.

Calcification occurs when VSMCs are cultured in the medium containing high phosphate (calcification medium) (4, 25, 41). When incubated in the calcification medium, the senescent VSMCs demonstrated dramatically enhanced calcification compared with the control young cells (Fig. 2A). Since apoptosis has been reported to contribute to the progression of VSMC calcification in vitro (35), we explored whether apoptosis was increased in the senescent VSMCs. The number of apoptotic cells was not significantly different between the senescent and the young VSMCs cultured in the calcification medium, suggesting that a factor(s) other than apoptosis causes the enhanced calcification in the senescent VSMCs (Fig. 2B).

Enhanced expression of osteoblastic genes in the senescent VSMCs.

Since bone calcification regulatory factors locally expressed in blood vessels have been suggested to play significant roles in the formation of vascular calcification, we investigated the gene expression of these factors in the senescent VSMCs. Expression of MGP, the best known anti-calcification factor in blood vessels, was significantly downregulated in the senescent VSMCs compared with that in the control young cells (Fig. 3A). On the other hand, the expression of genes highly expressed in osteoblasts, such as ALP and type I collagen but not osteocalcin, was markedly enhanced in the senescent VSMCs (Fig. 3A). In contrast, the expression of type II collagen, the major ECM of cartilage that is highly expressed in chondrocytes (1, 26, 31, 39), was not different between the senescent and the control young VSMCs (Fig. 3A). Expression of bone morphogenetic protein-2 (BMP-2) and the sodium-dependent phosphate cotransporter Pit-1 was not altered in the senescent VSMCs (Fig. 3A).

We also examined the effect of high phosphate on the expression of these genes in the senescent VSMCs. High phosphate further enhanced the expression of type I collagen and tended to increase ALP expression, whereas it did not affect the expression of RUNX-2 and Pit-1 in the senescent VSMCs (Fig. 3B).

Because BMP-2 potently initiates osteoblastic differentiation and MGP negatively regulates BMP-2 signaling, we investigated the BMP-2 signaling in the senescent VSMCs. Phosphorylation of Smad 1/5/8, downstream signal molecules for BMP-2, was rather decreased in the senescent VSMCs (Fig. 3C). This may suggest that once the senescent VSMCs acquire the osteoblast-like phenotype, the sensitivity to BMP-2 might be reduced. However, the contribution of BMP-2 signaling in the process of osteoblastic transition during the senescence remains to be elucidated.

Inconsistent with the increased mRNA level, ALP enzymatic activities were also enhanced in the senescent VSMCs (Fig. 3, D and E). These results suggest that VSMCs undergo the osteoblastic transition during the senescence. This senescence-mediated osteoblastic transition in VSMCs are well consistent with the previous findings that MGP expression was reduced and that ALP expression was enhanced in the human vessels calcified at the media in vivo (40).

Increased expression of ALP and type I collagen enhances VSMC calcification in the senescent cells.

We then explored whether increased expression of ALP and type I collagen is responsible for the enhanced VSMC calcification in the senescent cells. Knockdown of either ALP or type I collagen significantly reduced the calcification in the senescent VSMCs (Fig. 4A). Knockdown of target genes was confirmed by the quantitative PCR analysis (Fig. 4B). These results indicate that increased expression of osteoblastic genes directory affects the susceptibility to calcification in the senescent VSMCs.

RUNX-2 has been reported to play a central role in the osteoblastic differentiation as well as calcification of VSMCs (4, 6, 30, 43). Therefore, we investigated the possible role of RUNX-2 in the osteoblastic transition of VSMCs during the replicative senescence. RUNX-2 expression was significantly upregulated in the senescent VSMCs, and knockdown of RUNX-2 significantly reduced the ALP expression and increased the MGP expression in the senescent VSMCs (Fig. 4, C and D). Furthermore, RUNX-2 knockdown significantly reduced the VSMC calcification in the senescent cells (Fig. 4E). These results indicate that RUNX-2 plays a role in the osteoblastic transition and the enhanced calcification in the senescent VSMCs. However, the expressions of ALP and MGP were not fully restored, and RUNX-2 knockdown had no effect on the type I collagen expression in the senescent VSMCs (Fig. 4D). Therefore, RUNX-2 appears not to entirely regulate the osteoblastic transition in VSMCs during the senescence.

On the other hand, other transcriptional factors critically involved in the osteoblastic differentiation, such as Msx-2 and osterix (23), were not increased in the senescent VSMCs (Fig. 4C). Expression of osterix was rather decreased in the senescent VSMCs compared with that in the control young cells. Thus osteoblastic transition of VSMCs during the senescence is likely distinct from the physiological osteoblastic differentiation.

In vivo emergence of osteoblast-like cells in the calcified media.

To show the in vivo association between the cellular senescence and medial calcification, we analyzed the calcified aorta from an aging model mouse, klotho^−/− mouse. The klotho knockout mouse displays many age-related phenotypic characteristics, including short lifespan, osteoporosis, emphysema, and vascular medial calcification (25). An anti-aging function of the klotho gene has been reported in endothelial cells (20), and we found that VSMCs, rather than endothelial cells, highly expressed the klotho gene (Fig. 5A). Therefore, we investigated whether the osteoblastic transition of VSMCs occurs in vivo by using the klotho^−/− mice. Aorta from klotho^−/− mice demonstrated significant medial calcification as detected by von Kossa staining (Fig. 5B). Of note, significant expression of RUNX-2 was observed exclusively in the calcified media (Fig. 5B). These findings further support our hypothesis that osteoblastic transition in the senescent VSMCs plays a crucial role in the medial calcification associated with aging. On the other hand, MGP expression was detected at the boundary region of medial calcification as previously reported in human calcified plaque (9).

Statin has protective effects on P_i-induced apoptosis and calcification in the senescent VSMCs.

Statin treatment was recently shown, with the use of VSMCs in early passage, to inhibit the VSMC calcification by inhibiting the P_i-induced apoptosis (41, 42). However, it remains to be elucidated whether statin also has a protective effect on the P_i-induced apoptosis and calcification in senescent VSMCs. In addition, the effect of statin on the gene expression of bone calcification regulatory factors is not known in VSMCs.

Fluvastatin significantly reduced the calcification in the senescent VSMCs as well as in the control young cells (Fig. 6A). P_i-induced apoptosis was also dramatically reduced by fluvastatin in both the senescent and the young VSMCs (Fig. 6B). Because inhibition of Rho-kinase is attributed to the pleiotropic effects of statins (36), we examined the effect of Rho-kinase inhibitor Y-27632 on the P_i-induced apoptosis and calcification in VSMCs. Rho-kinase inhibitor significantly reduced the P_i-induced apoptosis and calcification in the senescent VSMCs as well as fluvastatin (Fig. 6, A and B). In young VSMCs, Rho-kinase inhibitor significantly reduced the P_i-induced apoptosis and also tended to reduce the calcification. These results suggest that the inhibition of Rho-kinase is possibly involved in the protective effect of statin against the P_i-induced apoptosis and VSMC calcification, at least in the senescent VSMCs.

Treatment with fluvastatin or Rho-kinase inhibitor significantly increased MGP expression in both the senescent and the young VSMCs (Fig. 6, D and E). Considering that MGP is a potent calcification inhibitor in blood vessels, it is possible that this increase of MGP is involved in the mechanisms through which statin and Rho-kinase inhibitor inhibit the VSMC calcification. On the other hand, neither fluvastatin nor Rho-kinase inhibitor significantly affected the expression and activity of BMP-2 as assessed by quantitative PCR and immunoblotting for the phosphorylated Smad 1/5/8, respectively (Fig. 6, D and E).

MGP has five γ-calboxyglutamic acid (Gla) residues that are essential for its anticalcification function (33, 34). On the other hand, Gas6, which plays a crucial role in the statin-mediated antiapoptotic effect (41), also contains Gla residues essentially involved in its survival signaling (15). Because the production of Gla residues is dependent on vitamin K, warfarin is known to inhibit the function of both MGP and Gas6. We therefore examined the effect of warfarin on the anticalcification effect of statin and Rho-kinase inhibitor. Treatment with warfarin completely abolished the beneficial effect of statin and Rho-kinase inhibitor on the VSMC calcification (Fig. 6A), suggesting a role of Gas6 survival signal and MGP in the anticalcification effect of statin and Rho-kinase inhibitor. However, we cannot exclude the possibility that another factor(s) is involved in the anticalcification effect of statin and Rho-kinase inhibitor, because warfarin is not a specific inhibitor for Gas6 and MGP, and it is not clear which of the two proteins is important in mediating the anticalcification effect.

Treatment with warfarin further enhanced the calcification, even compared with the vehicle control, only in the young VSMCs (Fig. 6A). Consistent with this finding, treatment with warfarin deteriorated VSMC calcification only in the young VSMCs, not in the senescent VSMCs (Fig. 6C). This different effect of warfarin is probably due to the loss of MGP in the senescent VSMCs; MGP expression is too low to effectively inhibit the VSMC calcification, and thus inhibition of MGP may not affect the calcification in the senescent VSMCs.

These data suggest that statins and Rho-kinase inhibitor might be useful for the prevention as well as the treatment of the senescence-associated medial calcification.

DISCUSSION

Medial artery calcification is well known to be associated with senescence, but the molecular link between its formation and cellular senescence remains poorly understood. Although the potential association among the senescence, osteoblastic transition of VSMCs, and medial calcification has been suggested by the findings that VSMCs in the human calcified media express osteoblastic genes in vivo (40), their correlation has never been demonstrated clearly both in vitro and in vivo. In the present study, we showed that VSMCs underwent the osteoblastic transition during the replicative senescence at least partially through a RUNX-2-dependent mechanism and that this osteoblastic transition directory led the higher susceptibility to VSMC calcification. Thus cellular senescence in VSMCs potentially plays a pivotal role in the formation of medial calcification associated with senescence.

Cellular senescence is the limited ability of primary human cells to divide in vitro, accompanied by specific phenotypic changes in morphology, gene expression, and function (12, 28). These phenotypic changes probably relate to age-related diseases (12). In the present study, we used the replicative senescent VSMCs as an in vitro model to study the role of senescent VSMCs in the vascular calcification. Because VSMCs in the tunica media do not actively proliferate in vivo, it is arguable whether replicative senescent VSMCs in vitro well reflect the senescent VSMCs in vivo. However, it has been reported that VSMCs in tunica media of old rats demonstrate senescence markers such as senescence-associated β-galactosidase activity and increased p16^Ink4a expression in vivo (47), which were also detected in the replicative senescent VSMCs in vitro. Furthermore, VSMCs of individuals whose life is near the end have been reported to show extremely low proliferation rates in the culture, which is also a unique feature for the replicative senescent cells (37). Therefore, we believe that replicative senescent VSMCs are plausible model to study the role of senescent VSMCs in vitro.

A detailed molecular mechanism responsible for the osteoblastic transition in the senescent VSMCs remains to be elucidated. RUNX-2 is a core transcriptional factor for osteoblastic differentiation of precursor cells (21, 23) and plays an essential role in the osteoblastic differentiation of VSMCs induced by hydrogen peroxide (4) as well as by uremic serum (6). Our current results demonstrate that RUNX-2 also contributes to the osteoblastic transition in the senescent VSMC, yet a RUNX-2-independent mechanism also appears to exist. Given that RUNX-2 knockdown had no effect on the type I collagen expression in the senescent VSMCs, a factor(s) that regulates type I collagen expression should be involved in the RUNX-2-independent pathway in the osteoblastic transition of VSMCs during the senescence.

We observed that the expression of osteocalcin, a marker for mature osteoblasts (44), was not enhanced in the senescent VSMCs. This might indicate that the senescent VSMCs is not fully differentiated into the mature osteoblasts and might be associated with the reduced expression of osterix, because osteocalcin expression is reported to be synchronous with the osterix expression (17). Therefore, replicative senescent VSMCs are the osteoblast-like cells but not the bona fide osteoblasts, and the osteoblastic transition during the senescence is likely a unique process different from the physiological osteoblastic differentiation.

Genetic mutation of klotho causes multiple premature aging-like phenotypes, and shortens lifespan down to 5–6% of that in wild-type mice (46). These aging-like phenotypes are largely attributable to the excess activity of 1,25-dihydroxyvitamin D₃ due to impaired fibroblast growth factor-23 signaling (46). Nevertheless, klotho gene is indeed involved in the process of cellular senescence, because overexpression of klotho markedly reduced the cellular senescence both in vitro and in vivo (14, 20). Furthermore, it has been reported that high concentration of plasma phosphate is not sufficient to induce vascular calcification (32). Therefore, we believe that the klotho^−/− mouse is the appropriate model to investigate a link between the cellular senescence and vascular calcification in vivo, and our observation that osteoblast-like VSMCs emerge in the calcified media in klotho^−/− mice strongly supports our hypothesis that osteoblastic transition in VSMCs during the senescence plays a crucial role in the senescence-associated vascular calcification.

We demonstrated that statin and Rho-kinase inhibitor prevent the VSMC calcification in both the senescent and the control young VSMCs, presumably by reducing the P_i-induced apoptosis and increasing the MGP expression. These results suggest that statin and Rho-kinase inhibitor might be feasible to prevent the senescence-associated medial calcification. Although the clinical use of statins is unlikely very effective to inhibit the vascular and valvular calcification (5, 7, 13, 38), the discrepancy between the in vitro and in vivo outcomes could be explained by 1) statins might have a beneficial effect on the inhibition of initial vascular calcification rather than on the progression of established calcified lesions, and 2) different statins might have different pleiotropic effect on the vascular and valvular calcification.

Effective therapies to treat and prevent the vascular calcification have not yet been available on the clinical setting. Our findings suggest that osteoblastic transition of VSMCs during the senescence would be a potential pharmacotherapeutic target to prevent the senescence-associated medial calcification.

GRANTS

This work was supported by research grants from the Ministry of Education, Culture, Sports, Science and Technology, Japan (KAKENHI 18790507 and 20590885), a Sakakibara Memorial Research Grant from the Japan Research Promotion Society for Cardiovascular Diseases, the Takeda Science Foundation, and a Japan Heart Foundation/Novartis Research Award for Molecular and Cellular Cardiology.